TjongKristian, yes, of course it's good to treat people with respect. However, this is a discussion forum where people are I think above all trying to learn. Yes, replies to erin's posts have been sceptical, ironic and sometimes a little provocative, but not at all amazing, particularly given the nature of his claims.
Further, those claims have been posted up here without ABX testing having been done on them. When Pano suggested free software to do, this erin's response was dismissive and, in my opinion, quite rude.
Quite understandably, there are people on here, and I'm probably one of them, who are now rather beyond tired of doo-lally subjectivist BS. I think I've read all the replies posted thus far, and I don't think there's a single one of them that's beyond the pale.
Second that.
TjongKristian, if the files are exactly the same bitwise, they are exactly the same, there is no difference between them, so if they sound different it is due to some fault of the playback system. It is the claim that the files are the same, yet sound different due to the ripping process that is being disputed, and it is this claim that is being discussed.
As I philpoole and others, myself included have said, its data, nothing more, to the disk operating system, its bits that are moved from one place to another with all the other data that is transferred around a system, but on the disk it is just data a 1 or a 0.
TjongKristian, if the files are exactly the same bitwise, they are exactly the same, there is no difference between them, so if they sound different it is due to some fault of the playback system. It is the claim that the files are the same, yet sound different due to the ripping process that is being disputed, and it is this claim that is being discussed.
As I philpoole and others, myself included have said, its data, nothing more, to the disk operating system, its bits that are moved from one place to another with all the other data that is transferred around a system, but on the disk it is just data a 1 or a 0.
It is irrelevant. The data comes off the disk the way it is written, bit perfect.
Hard disks read/write data in blocks. There is a read/write buffer that holds chunks of data so that the process that reads/writes the data, whether it is an audio playback or loading the picture editing plug-in for Word, is served by data when it is requested. This can all be asynchronous, but the DAC that needs the data to make it into analog gets it's data bit perfect and on time. No surprises here whether the disk is defragmented or whatever.
jan
Third that.
If Erin was looking for the truth he would do the file comparison, so obviously hes not. Some people cant admit there wrong, so let him waste his time re-ripping his library.
Non of this makes any difference! The data is read of the HDD as fast as possible (probably 100x the speed of the audio) and put into a RAM buffer (these days large enough to put the whole CD in if the playback software wants to) where it is clocked out by the PC processor or outboard DAC. The HDD has nothing to do with the timing (jitter). (silver sata cable sounds better my ***).
The data comes of the HDD bit perfect, wave files have the necessary error correction. When was the last time a PDF came up with the wrong letter in a word?
If Erin was looking for the truth he would do the file comparison, so obviously hes not. Some people cant admit there wrong, so let him waste his time re-ripping his library.
Non of this makes any difference! The data is read of the HDD as fast as possible (probably 100x the speed of the audio) and put into a RAM buffer (these days large enough to put the whole CD in if the playback software wants to) where it is clocked out by the PC processor or outboard DAC. The HDD has nothing to do with the timing (jitter). (silver sata cable sounds better my ***).
The data comes of the HDD bit perfect, wave files have the necessary error correction. When was the last time a PDF came up with the wrong letter in a word?
Last edited:
There's been much discussion so far, but I don't think this was asked: What mode were the files played back in? Was there any difference in playback between regular mode and safe mode?
I hope QUESTIONS are okay...
It took me a monent to find that, as the spelling is fidelizer, not fedelizer:
Windows X's Live | Fidelizer
Looks like something that's trying to fix things that ain't broke. It might have been useful (depending on what it actually does, which I didn't quite get from that description) back in the 200MHz Pentium days, but back then I knew what processes to disable so audio wouldn't stutter due to soundcard buffers underflowing.
Actually a one-bit difference can be heard, especially in a quiet passage and if it's one of the more significant bits that gets flipped. It sounds remarkably like a tick on an LP.
This is an interesting attitude. If I found a technique or process that makes something sound better (or worse, or different in any way) that I "knew" shouldn't have any effect, I'd be investigating and asking myself AND DIYaudio "What could be happening here? I swear this sounds better to me, but why?" rather than re-ripping every CD I have.
It won't shut me up, but it will certainly make me raise an eyebrow.
It's not JUST that it requires work on our parts, but even moreso it's an assertion many of us find incredulous, much like a previous thread claiming that playback from a flash memory stick sounds different from playback from hard drive of an exact copy of the same music file. I recall that thread became so cantankerous that it eventually got locked.
I think it was Carl Sagan who said "Extraordinary claims require extraordinary evidence" and the evidence should be presented by the person making the claim.
@Wombat
I wouldn't mind so much, but that "kind of people" have, collectively, monopolised the hobby of high-quality music production.
In the Audio Critic's last issue in their online archive (dated some time in the 90s if memory serves), they explain that they are giving up with their "Hip Boots" column, which highlighted major pieces of BS/technical incompetence on the part of hi-fi magazines, for the simple reason that the entire audio press appears to have gone insane.
@Others
Why exactly are we trying to DBT this? If I declared that changing my wallpaper to a picture of a kitten altered my system's sound quality should we all leap up and immediately start testing? Last I checked, if I email the physicists at CERN and make a completely unsubstantiated assertion - say, that they could find the Higg's Boson if they would only fill their particle accelerators with cake - they're not all going to sprint down to their local bakery in the interests of science.
I wouldn't mind so much, but that "kind of people" have, collectively, monopolised the hobby of high-quality music production.
In the Audio Critic's last issue in their online archive (dated some time in the 90s if memory serves), they explain that they are giving up with their "Hip Boots" column, which highlighted major pieces of BS/technical incompetence on the part of hi-fi magazines, for the simple reason that the entire audio press appears to have gone insane.
@Others
Why exactly are we trying to DBT this? If I declared that changing my wallpaper to a picture of a kitten altered my system's sound quality should we all leap up and immediately start testing? Last I checked, if I email the physicists at CERN and make a completely unsubstantiated assertion - say, that they could find the Higg's Boson if they would only fill their particle accelerators with cake - they're not all going to sprint down to their local bakery in the interests of science.
Last edited:
Absolutely.
And so far we've not been presented with any evidence at all. Just opinion, and instruction that we must gather the evidence ourselves.
Suggestions on easy ways to check some basics (like does this stand up to ABX testing - surely the obvious and necessary test) have been rejected or ignored.
It's been a while since I dug into USB DACs. But last I did, most were linked to clocking done on the PC side. Most of the USB audio receiver chips did *not* have an asynchronous receiver mode (with memory and independent crystal type clocks on board). The reference used to clock the USB DAC was received from the PC.
The interesting part of this, was at the time, most people in that forum discussion preferred the sound of these same USB DACs. Go figure.
There's fundamental differences between streaming and "copying" digital material. When streaming audio or video in realtime, the timeliness of delivery and quick recovery of time sync after errors are more important than completeness, and the protocol and prorities reflect that. So if you're streaming, there may be value to killing those extra task and watching u-tube later and hoping your neighbor's kid isn't on some webcam. Streaming transmission assumes a very short dropout is no big deal but jump in time are a big problem.
But the wonderful thing of good digital encoding (and what makes it fundamentally different from analog copy processes) is the assurance that a copy (assuming competent redundancy, error detecdtion, and error correction protocol) can be guaranteed 100% perfect (with a confidence level that means you won't see an error in your lifetime).
Machines have a hard time tying to decide how to best "fake it" when there is a long fatal media flaw in your CD, but when you copy a digital file and check it, it's either intact of it's not...there are no other "quality" issues like with an analog copy. An analog copy, no matter how good, is slightly altered every time it's copied. But with proper digital copies, if each copy is successful, then the 1 billionth generation copy of a copy of a copy of a copy etc. will still be perfect...or the operation aborted. All or nothing is the general rule for copying digital data!
But the wonderful thing of good digital encoding (and what makes it fundamentally different from analog copy processes) is the assurance that a copy (assuming competent redundancy, error detecdtion, and error correction protocol) can be guaranteed 100% perfect (with a confidence level that means you won't see an error in your lifetime).
Machines have a hard time tying to decide how to best "fake it" when there is a long fatal media flaw in your CD, but when you copy a digital file and check it, it's either intact of it's not...there are no other "quality" issues like with an analog copy. An analog copy, no matter how good, is slightly altered every time it's copied. But with proper digital copies, if each copy is successful, then the 1 billionth generation copy of a copy of a copy of a copy etc. will still be perfect...or the operation aborted. All or nothing is the general rule for copying digital data!
Steve, if you're feeling bored and/or ubergeeky:
http://en.wikipedia.org/wiki/Fisher's_exact_test
Originally devised to test the claim of a woman that she could tell whether the milk had been added to her cup before or after the tea. She got all eight right, supposedly...
I make the probability of getting all 10 correct by random guesswork to be about 0.4% - 1% is normally regarded as "highly significant" statistically speaking - (SY, hope my stats looks right to you).
If Erin can correctly identify all 10, a commotion will indeed ensue, and rightly so! (I'm almost rooting for him now, how odd!)
http://en.wikipedia.org/wiki/Fisher's_exact_test
Originally devised to test the claim of a woman that she could tell whether the milk had been added to her cup before or after the tea. She got all eight right, supposedly...
I make the probability of getting all 10 correct by random guesswork to be about 0.4% - 1% is normally regarded as "highly significant" statistically speaking - (SY, hope my stats looks right to you).
If Erin can correctly identify all 10, a commotion will indeed ensue, and rightly so! (I'm almost rooting for him now, how odd!)
Perhaps rip a CD on a laptop in safe mode & runs off battery will sounds even better
Hmmm, it would mean you could do it inside a Faraday cage.....
When I worked for a certain large mid-fi audio company back in the dark ages we considered 70% or greater success statistically significant in A/B/X testing, enough that we paid serious attention to anyone who achieved these results in evaluating the acoustic performance of our designs. The interesting thing is the guys who mixed and mastered the demo CDs provided with the various products consistently had very high scores in ABX testing with completely unfamiliar hardware and in many instances where others often failed to hear significant differences.
Nah, the extra power provided by the power supply makes the read sensor more sensitive and the motor more smooth.
And I don't understand why it would sound better and not worse - from my experience with Windows from 95 to XP, DMA will be turned off in safe mode (my Win 7 system hasn't crashed enough to required booting in safe mode) and the system reverts to the old PIO data transfer method, and apart from making safe mode load very slowly, it will also require data to pass through the CPU and meeting dirty electrons. (In comparison, with DMA enabled the data goes straight from drive to memory.)
Perhaps that "better" is due to dirty electrons?
And I don't understand why it would sound better and not worse - from my experience with Windows from 95 to XP, DMA will be turned off in safe mode (my Win 7 system hasn't crashed enough to required booting in safe mode) and the system reverts to the old PIO data transfer method, and apart from making safe mode load very slowly, it will also require data to pass through the CPU and meeting dirty electrons. (In comparison, with DMA enabled the data goes straight from drive to memory.)
Perhaps that "better" is due to dirty electrons?
Perhaps rip a CD on a laptop in safe mode & runs off battery will sounds even better
Many people even run laptops off battery for serious listening sessions.
Last edited:
Well, the significance would depend on how many tests were involved, and whether the subjects knew how many As and how many Bs there were going to be. The more tests, the more significant any given percentage correct becomes.
- Status
- Not open for further replies.